Friday 19 May, 2006

How to start antidepressants

Posted in General Practice, May Su, Psychiatry, Wards at 23:12 by May Su

Original article by: May Su :: Printer friendly

The antidepressants that would be used as a first line agent would usually be a selective serotonin re-uptake inhibitor (SSRI), venlafaxine or mirtazapine. The choice would be determined by the symptoms of the patients and the side-effect profile that would be tolerable. At times, a side effect of a medication may be used as an advantage. See the antidepressant matrix for a single page overview. Furthermore see the list of antidepressants available in Australia.

Of “first line” agents, a few generalisations can be made:

SSRIs, venlafaxine and mirtazapine:

  • The base effective dose is equivalent to one tablet daily;
  • “start low, go slow”: start with a half tablet daily for 4-6 days (except fluoxetine which has very long half life and can be started as one tablet daily);
  • then increase to one tablet daily;
  • the dose can be increased to at least 2 tablets daily (though there is no “rule of thumb” of the maximum safe daily dose – each drug is different).

Before starting a patient on an antidepressant, discuss side-effects. The most common side effects are nausea, headache, sexual dysfunction, agitation and sleep disturbance. It is useful to emphasise that early side-effects like agitation, sleep disturbance and nausea often settle after the first few days to week of treatment and it is important to press on.

The principle is to “start low, go slow”. Start at “half dose” for 4 days and increase to the full dose afterwards. For elderly patients, it may be appropriate to continue on the “half dose” until review. Warn patients that this is a long term medication, usually of a minimum of 18-24 months for a first episode.

Initially patients should be reviewed weekly or fortnightly. This is to assess side effect profile, continuing risk and for supportive therapy. Patients need to be aware that it usually takes 4-6 weeks before they receive a reasonable effect from antidepressants, although there should be some effect by two weeks of therapy (3-4 weeks for fluoxetine due to long half life). If there is no noticeable effect by 2 weeks, the dose should be increased (increase the dose by 50-100%). If response is unsatisfactory at 6 weeks, increase the dose again. If there is no effect at all at the 6 week mark (despite having increased the dose), it may be worthwhile to change antidepressants. Otherwise review progress at 4 weeks after each dose adjustment.

The most common reason for poor response to antidepressants is not using a reasonably high dose, or not trialing for long enough. There is some evidence that fluoxetine (and hence SSRIs) work at the level of neurogenesis (1). They don’t work quickly.

There is a risk of relapse if precipitously ceasing antidepressants. Suddenly stopping antidepressants can also cause significant withdrawal effects, so weaning over several weeks is the preferred method.

Antidepressants do not necessarily have to be continued indefinitely. If patients have been symptom free, then it is not unreasonable to consider a trial off antidepressant after 18-24 months, however they must be warned of risk of relapse and withdrawal.

SSRIs are highly effective. There has been some bad press in the past few years related to increased risk of suicide with SSRIs. The main reason for this is that SSRIs (in fact all antidepressants) take longer to improve mood than they take to improve the other symptoms of depression such as amotivation. This is an issue for the severely depressed patient who has had significant suicidal ideations, or even plans, but mainly lacks the concentration and interest to carry out the attempt. As the antidepressants act, the patient’s functional ability improves while they may still have symptoms of nihilism and low mood. This is when they require closer monitoring regarding risk to self or others. A study showed that patients with depression on SSRIs had around twice the number of suicide attempts compared to those on placebo, though there was no statistical significance for fatal attempts. Furthermore, there was no statistical significance between SSRIs and TCAs (though paroxetine was on the edge of significance).

Choice of antidepressants – by side-effects:

Avoidance of drug interactions:

  • citalopram
  • sertraline
  • venlafaxine
  • mirtazapine
  • avoid: fluoxetine

Avoidance of sedation:

  • any of the SSRIs except fluvoxamine
  • venlafaxine
  • avoid: mirtazapine

Patient with insomnia (sedation may be a useful side-effect):

  • mirtazapine
  • fluvoxamine is relatively sedating for an SSRI

Patient with severe anxiety with their depression:

  • mirtazapine
  • avoid: venlafaxine

Patients with severe depression:

  • venlafaxine and;
  • mirtazapine may be “stronger” than the SSRIs.
  • avoid: moclobemide

Avoidance of sexual side-effects:

  • moclobemide
  • mirtazapine

Choice of antidepressants – by patient group

Elderly patient with polypharmacy:

  • citalopram
  • sertraline
  • mirtazapine

Otherwise healthy adult:

  • All the SSRIs are usually well tolerated;
  • sedation and weight gain may be an issue with mirtazapine.

Pregnant or breastfeeding woman:

  • Best evidence of safety with TCAs and fluoxetine;
  • most SSRIs are probably safe;
  • there are lower levels of sertraline and paroxetine in breast milk compared to fluoxetine but this is of unknown significance.
  • avoid in pregnancy: paroxetine (recent research showing a statistically significant increase in cardiovascular malformations – see more below) (2) (3).

Sexual side-effects:

  • Can try to switch to another SSRI;
  • mirtazapine has no sexual side-effects but sedation and weight gain may be issues;
  • moclobemide has no sexual side-effects by relatively less effective.

In adolescents and children:

  • Poor evidence base for most medications (specialist referral recommended);
  • fluoxetine.

Reference links:

(1) Huang GJ., Herbert J. Stimulation of neurogenesis in the hippocampus of the adult rat by fluoxetine requires rhythmic change in corticosterone. Biol Psychiatry. 2006 Apr 1;59(7):619-24. Epub 2005 Dec 2.

(2) Therapeutic Goods Administration (TGA). Caution over antidepressant paroxetine during pregnancy [media release]. 7 September 2005. [download PDF :: 30 Kb]

(3) GlaxoSmithKline. Paroxetine and pregnancy [website].

Updated: Michael Tam (19 June 2003)

Please read the disclaimer


  1. Michael Tam said,

    Advised by Dr May Su to change recommendations on antidepressants to use in pregnancy to include recent epidemiological evidence suggesting potential harm from paroxetine.

    From the GlaxoSmithKline website page on "Paroxetine and pregnancy" (article reference (3)):

    "GlaxoSmithKline sponsored a retrospective, U.S. epidemiologic study of major malformations following maternal exposure to antidepressants in the first trimester. This analysis showed a trend towards a 1.5-fold increased risk for cardiovascular malformations for paroxetine compared to other antidepressants. This study also showed a statistically significant increased overall risk of major congenital malformations (inclusive of the cardiovascular defects) in infants exposed to paroxetine compared to other antidepressants. A second, independent retrospective epidemiological analysis utilizing the Swedish national registry database reported a 2-fold increased risk of cardiac defects in infants exposed to paroxetine, compared with the general population."

    As such, paroxetine is not recommended for use in pregnancy.

  2. Michael Tam said,

    A recent study (1), however, has not been able to confirm the effect.

    Wolfgang Paulus at the Institute of Reproductive Toxicology in Ulm, Germany, looked at the pregnancy outcomes of 119 women who took paroxetine during the first trimester of pregnancy and compared them to 645 women who had not been exposed to the drug.

    “We found that the rate of congenital abnormalities was not increased after using paroxetine in early pregnancy,” says Paulus. 

    The facts are still out there but given that there are many other SSRIs to choose from, it would seem prudent to avoid paroxetine in early pregnancy.


    (1) Geddes, L.  Common antidepressant does not increase birth defects [electronic article]. news service. 19 June 2006

  3. Gibran said,

    Women that are considering taking antidepressants should research the the medication they are considering to take, and women that are pregnant should seek professional medical advice before beginning any treatment for depression. If you are considering taking Paxil you may be interested in reading more about Paxil (paroxetine) birth defects.

    URLs removed by Administrator: You were treading a fine line with this post between advertising your legal services and providing information. As such, I have removed your links. Furthermore, your website is hardly a medical or scientific authority on the increased risk of congenital abnormalities and paroxetine. Please see the aforementioned links in this thread.

  4. Michael Tam said,

    Thank you for the American perspective.

    I entirely agree; anyone (pregnant or not) should seek professional advice before beginning any treatment for depression. In Australia though, medications are not advertised to the consumer and it would be impossible to commence a pharmaceutical like an antidepressant without first having seen a doctor.


  5. med student said,

    Hi, I’m a 6th year med student studying for my exams – thanks for this, it was incredibly helpful! As is the rest of your site!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: